Data network topology

Instead of software based problems, this thesis focuses more on hardware and routing related challenges. SRtP rules state that the ship’s automation system must remain oper-ational in any fire or flooding casualty scenario. To secure a fully SRtP redundant sys-tem, there must be emphasis on the importance of the data LAN topology and how it is designed, as the SRtP scenarios are formed with a space-by-space related approach.

This thesis examines the possibility of using IAS as the basis for a complete data network. It is an existing system, distributed around the ship, and already equipped with easy connection interfaces to other systems (with normal suppliers). In the reference vessel however, IAS process stations are quite centralized and I/O cabinets do not posses networking capabilities (each I/O cabinet is automatically connected to both process stations). This could cause problems, or additional installation costs, if sub-systems are connected as part of the network. One solution is to further distribute the system, using more sophisticated RTUs with network capabilities and sophisticated connections (such as Ethernet TCP/IP). A more distributed and sophisticated IAS would provide more layout possibilities, system security (from SRtP point-of-view) and possi-bly lower cabling costs. Yet, even if savings with reduced cable costs could be acquired, the more developed IAS would most likely increase acquisition costs. Economical as-pect must be kept in mind when designing the system. However, economical asas-pects should not be the only designing factor, especially if system integration creates new possibilities and possible savings from other areas. There is a possibility however, that a larger network could cause problems in the future. These problems may be related with possible real-time processes working under IAS, and with network message quality, collisions and response times, depending on the network topology and used solutions.

In this case, the scope for the assessment of the system network is defined not to ex-amine connections to sub-systems and is instead defined to only exex-amine IAS with its connections. According to the delivery specification, the reference ship’s IAS consists of two process stations, a Safety Console and 7 I/O cabinets. Two main operating sta-tions are provided – one at the bridge and one in ECR. One reserve operating station is located at the after switchboard room. Safety Console and process stations form a logi-cal dual ring network and all I/O cabinets have duplicated connections to both process stations (see Appendix 1). IAS is connected to following separate systems with serial line connections: Integrated Navigation System, Voyage Data Recorder, Remote Gaug-ing System, LoadGaug-ing Computer, Remote Controlled Valves, Main Engines (4) and Aux-iliary Engines (4) and to Propulsion Power board System. [2.]

Developing a functional concept for IAS network topology is challenging. Accord-ing to Kenyon, designAccord-ing a low cost, yet efficient, network topology relies on very complicated algorithms and heuristic techniques. Depending on the amount of locations (N), there is × ( 1)/2 possible links and 2 ^{×( )/} possible topologies for a network. [23.] For example, with 7 locations there are 2 097 152 possible topology op-tions. This means that there is no easy way to secure the best possible option. All design processes depend on the amount of locations and requirements, called constrains. Ke-nyon continues that a design process starts with supporting data, gathered during capaci-ty planning phase, and a designer should try to satisfy certain chosen constrains, such as reliability, delay-throughput and costs. [23.]

With a ship’s data network, one constrain should be added - redundancy. This, how-ever, reduces the amount of possible topologies only slightly. If less sophisticated I/O stations are used, it reduces the amount of topologies but increases the amount of cabl-ing. Sophisticated I/O cabinets, with network options, would enable a possibility for a mixture of different network topologies, such as a logical two way ring (figure 6.2) used between the process stations, but would increase the acquisition costs.

Figure 6.2. Two way ring topology.

The advantage with the logical two way ring topology is that it automatically secures a redundant connection. Meaning, that in any cable failure scenario there already exist an alternative cable route. This means that the SRtP rules are already partially complied. In a cabinet failure scenario, a two way ring transforms into a C-topology and information is restored as long as the sensor/actuator is connected to at least two cabinets, it is lo-cated in the same space with the cabinet or one connection is acceptably secured (fire resistant cable). When I/O cabinets are located in separate spaces and the ring cabling is carried out as straightforward as possible (cables between different I/O cabinets not en-tering the same space), the system should survive all SRtP scenarios. Disadvantages could be problems with message quality and response time, geographical distances and possible cabinet layout problems (depending where I/O is needed). Also, it may not be beneficial to use a single ring to connect all cabinets but instead consider (depending on

the layout and distances) the use of hybrid topologies. These possibilities could be ex-ecuted with several field bus protocols, using switches to form a logical ring.

6.3 Propulsion system

The propulsion system satisfyingly fulfills both user and system requirements set by the operators and regulations. However, the design could be further developed from both economical and SRtP concept perspectives. The SRtP concept for propulsion system does not seem to have potential for any additional remote operated or automated fea-tures which it does not already posses. Failures are dealt with mechanical operations – using only one complete shaft line (shutting down the other) and/or manual valve opera-tions. Automated/remote controlled valve actions are, in most cases, not economically reasonable. Valves are scattered widely around the ship and in places easy to access.

However, the amount of manual actions could be reduced with changes in the design concept.

Two groups of valves should be examined more closely. First valve group consists of isolation and application valves for stern tube emergency tank and propeller hub gravity tank. These items are lost in scenarios where the RoRo deck of the ship is dam-aged. The operations for isolation/application cause a lot of actions on very theoretical basis and severely influence the capabilities of the ship if applied. According to Mr.

Todd, these features will hardly be used [11].

The use of gravity tanks depends on the supplier of the propellers. Gravity tanks are used to secure static pressure in the shaft line hub. As an alternative, there are suppliers who, instead of gravity tanks, offer pressure maintaining pumps. For the shipyard, gravity tanks have been used as a standard solution because it is easier to launch a ship into water with an incomplete shaft line (as, at that time there may not be working pow-er network on the ship to supply the pressure maintaining pump). A pump model could be both a more economical solution and it could also reduce needed actions in SRtP situations. With careful layout design the loss of the pump unit could be coupled with the loss of the complete shaft line. Hence, it does not add SRtP features, only reduces them. Instead of multiple valve actions, only a duplicated power supply with automatic changeover for the pump would be needed.

Second group of valves contain the isolation valves for instrument air. The use of instrument air for main engine rooms’ ventilation should be considered as an area for further development. In Spirit of Britain the fire dampers of both main engine rooms are pneumatically controlled, whereas the ventilation fans are electrically controlled. In SRtP sense, instrument air is only used for main engine room ventilation control. The control air is taken from instrument air system, which is connected to working- and starting air systems, provided from both main engine rooms and led to the funnel top via various decks and SRtP spaces. Additionally, the piping network for all of the com-pressed air systems spreads along the ship on multiple decks and spaces. A casualty anywhere on the piping network is considered to reduce pressure on the whole network.

Hence, there are a lot of isolation valve installations and analyzing work to be done for compressed air systems. This amount of work could be reduced greatly (or eliminated completely) with more sophisticated design concept.

There are at least two options how compressed air systems could be excluded from the SRtP scope. First is to change from pneumatic control to electrically controlled sys-tem. As this option would be easier from SRtP point-of-view, it would be more expen-sive and less reliable in emergency situations according to the project coordinator [18].

The second option is to change the design concept for main engine room fire dampers and ventilation fans. The current reference system is shown in figure 6.3. (Appendix 8).

In future vessels, new interpretations allow that the supply cables for ventilation fans could be installed from VMC (Ventilation Motor Control) centers up to the fans via main engine room and the funnel, provided that the cable is fire resistant. This would mean more expensive materials but reduced costs in working hours, as the route is shorter and easier for installation. This would decrease estimated total costs. Also, the pneumatic control boxes for the fire dampers could be installed on top of the funnels, on the weather deck, which is not part of any SRtP casualty scenario. Both boxes would be fed by a common instrument air vessel or an accumulator – also located on the weather deck. The vessel/accumulator would be connected to the instrument air system with two connections from both main engine rooms via engine room funnels, equipped with check valves. With this concept, the operation of both engine room dampers in every SRtP scenario is secured without any manual actions. As with the fan supply cables, the expenses for materials would increase but the total costs would reduce with saved work-ing hours and shorter routes.

However, there are certain criteria which must be met if the control boxes are si-tuated on the weather deck. First, the area containing the boxes must be restricted from all passengers to avoid any misuse of the equipment. Second, the boxes must be suitably protected (sufficient IP classification). And thirdly, the used instrument air must be dried according to prevailing weather conditions on the ship’s operating route to avoid unwanted freezing. The basic concept is shown in figure 6.4.

Figure 6.4. Main Engine Rooms’ ventilation concept.

From a more general point-of-view, the aim for modular propulsion system develop-ment should be that full propulsion and maneuvering capabilities (with two shaft lines) of the ship should remain operational in all but the most extensive fire/flooding scena-rios. The propulsion concept should be based on two shaft lines with four main engines in two main engine rooms (when reasonable/possible). The use of two shaft lines with four main engines allows redundancy both in SRtP scenarios and when maintenance work for propulsion system is needed. Basic design criteria should aim to reduce the amount of manual actions by design – meaning that the number of scenarios with a need of using only one shaft line would be reduced to a minimum, and manual actions is not needed except in the worst scenarios. This can be acquired with fire insulations on water cooling pipes, “A-60” rated trunks covering auxiliary systems as well as the shaft line, duplicated power supplies for auxiliary equipment and so forth.

The same concept as with fuel oil system could be applied: splitting the propulsion system into two independent circuits. When possible, a complete shaft line with all aux-iliary equipments would act as an independent system. For example, in Spirit of Britain the main engine rooms would function as an imaginary border for splitting cooling wa-ter and fuel oil systems. Systems for port side propulsion line would use systems si-tuated at spaces from forward main engine room to fore of the ship, and avoiding unne-cessary cross-connections to starboard side system. In casualty cases, it would always guarantee one fully working shaft line, with the minimum amount of actions, and possi-bly the other shaft line with reduced capabilities.